Keep pulling the thread on David Cahn.
The speaker, David, believes the technology industry is currently in an AI bubble.
The fragility of the current AI market is evident in the rise of circular financing deals and shifting capital expenditure dynamics among large tech companies.
Microsoft and Amazon have recently reduced their role in absorbing risk for the AI ecosystem's buildout, with Microsoft notably walking away from two data center projects.
Oracle and CoreWeave have significantly increased their role in supplying AI compute, stepping in as Microsoft and Amazon have pulled back.
The current AI buildout is primarily funded by equity and cash, not debt, meaning any market correction will likely be an equity unwind rather than a credit crisis similar to 2008.
The Magnificent Seven technology companies represent 40% of the S&P 500, creating a significant market concentration risk tied to the AI narrative.
The cost to build out one gigawatt of AI data center capacity is estimated to be $40 billion, with Jensen Huang of NVIDIA suggesting it could be $50-60 billion using next-generation Vera Rubin chips.
Leading AI labs like OpenAI and Anthropic are increasingly becoming vertically integrated by developing their own chips and procuring their own power.
The timeline for achieving Artificial General Intelligence (AGI) is being pushed out, with thought leaders like Andrej Karpathy, Richard Sutton, and Ilya Sutskever suggesting longer timeframes of 20-30 years.
The supply of industrial power generators is sold out until 2030, indicating a major physical constraint on data center construction.
The consensus view among tech leaders like Sam Altman, Vinod Khosla, and Jeff Bezos is that the AI sector is in a bubble.
Over 80% of venture capital investment in AI is currently directed towards producers of compute rather than consumers of compute.